Workflow Example

MaxQuant

For the analysis we require the proteinGroups.txt generated by MaxQuant and a corresponding metadata file where the sample names match with the sample names in the proteinGroups.txt file. At first take a look at the files:

import pandas as pd
import warnings
from plotly.offline import init_notebook_mode, iplot
warnings.filterwarnings("ignore", category=UserWarning, module="openpyxl") # remove warning from mac

The proteinGroups.txt contains all standard column headers from MaxQuant. Later, for our analysis we will use the Protein Intensity described in "LFQ intensity [sample]".

protein_groups = pd.read_csv("../../testfiles/maxquant/proteinGroups.txt", sep = "\t", low_memory=False)
protein_groups.head(5)

	Protein IDs	Majority protein IDs	Peptide counts (all)	Peptide counts (razor+unique)	Peptide counts (unique)	Protein names	Gene names	Fasta headers	Number of proteins	Peptides	...	Potential contaminant	id	Peptide IDs	Peptide is razor	Mod. peptide IDs	Evidence IDs	MS/MS IDs	Best MS/MS	Oxidation (M) site IDs	Oxidation (M) site positions
0	P01911;Q29830;Q9MXZ4;Q3LTJ8;Q3LTJ4;Q3LRY0;Q8HW...	P01911;Q29830;Q9MXZ4;Q3LTJ8;Q3LTJ4;Q3LRY0;Q8HW...	4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;...	4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;4;...	0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;...	HLA class II histocompatibility antigen, DRB1-...	HLA-DRB1;HLA-DR15;HLA-DRB1*;HLA-DRB1*1327;MHC ...	;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;...	1834.0	4.0	...	NaN	0.0	1287;1288;3971;15222	True;True;True;True	1415;1416;4387;16993	117340;117341;117342;117343;117344;117345;1173...	56384;56385;56386;159948;159949;159950;602426;...	56384;56385;159949;602427	NaN	NaN
1	P05121;A0A024QYT5;B7ZAB0;B7Z4X6;B7Z1D9	P05121;A0A024QYT5;B7ZAB0;B7Z4X6;B7Z1D9	10;10;9;8;5	10;10;9;8;5	10;10;9;8;5	Plasminogen activator inhibitor 1	SERPINE1	;;;;	5.0	10.0	...	NaN	1.0	1592;3771;4396;4628;6470;7404;9193;11222;13191...	True;True;True;True;True;True;True;True;True;True	1746;4166;4871;5128;7155;8188;10148;10149;1257...	136637;136638;136639;136640;136641;318672;3186...	64917;149749;149750;178775;178776;184126;25946...	64917;149749;178775;184126;259465;290355;36587...	NaN	NaN
2	P55083;A0A024QZ34;K7ES70	P55083;A0A024QZ34;K7ES70	2;2;2	2;2;2	2;2;2	Microfibril-associated glycoprotein 4	MFAP4	;;	3.0	2.0	...	NaN	2.0	140;15067	True;True	155;156;16823	13174;13175;13176;13177;13178;13179;13180;1318...	7599;7600;7601;7602;7603;7604;7605;7606;598026...	7604;598026	0	117
3	P09972;A0A024QZ64;A8MVZ9;B7Z3K9;B7Z1N6;B7Z3K7;...	P09972;A0A024QZ64;A8MVZ9;B7Z3K9;B7Z1N6;B7Z3K7;...	16;16;15;15;13;12;10;10;9;8;7;7;7;5	13;13;13;12;10;10;9;9;8;5;7;7;6;4	2;2;2;2;0;2;2;2;2;0;2;2;0;0	Fructose-bisphosphate aldolase C;Fructose-bisp...	ALDOC	;;;;;;;;;	14.0	16.0	...	NaN	3.0	312;749;1545;1675;2085;3265;5046;5708;8847;884...	True;True;True;False;True;True;True;True;True;...	346;827;1698;1838;2304;3606;5586;5587;6314;976...	29006;29007;29008;69091;69092;69093;69094;6909...	15520;15521;34013;34014;34015;34016;34017;6325...	15520;34015;63252;67942;84108;131214;198759;21...	1;2	40;251
4	Q96C19;H0Y4Y4;A0A024QZ77	Q96C19;H0Y4Y4;A0A024QZ77	1;1;1	1;1;1	1;1;1	EF-hand domain-containing protein D2	EFHD2	;;	3.0	1.0	...	NaN	4.0	8772	True	9685	734937;734938;734939;734940;734941;734942;734943	342134;342135	342134	NaN	NaN

5 rows × 2530 columns

Further, we created a excel-file with the corresponding metadata to our proteinGroups.txt-file. The sample names in the column “sample” match the names in proteinGroups.txt file.

metadata = pd.read_excel("../../testfiles/maxquant/metadata.xlsx")
metadata.head(5)

	subject external_id	biological_sample external_id	sample	tissue id	disease id	intervention id	tissue	disease	biological_sample quantity	biological_sample quantity_units	...	Alanine aminotransferase measurement (34608000)	Aspartate aminotransferase measurement (45896001)	Alkaline phosphatase measurement (88810008)	Gamma glutamyl transferase measurement (69480007)	Hemoglobin A1c measurement (43396009)	Total cholesterol:HDL ratio measurement (166842003)	High density lipoprotein measurement (17888004)	Low density lipoprotein cholesterol measurement (113079009)	VLDL cholesterol measurement (104585005)	Triglycerides measurement (14740000)
0	31	31	1_31_C6	BTO:0000131	NaN	NaN	blood plasma	healthy	NaN	NaN	...	24.0	30	54	21.0	6.3	3.6	1.26	2.1	0.3	0.58
1	32	32	1_32_C7	BTO:0000131	NaN	NaN	blood plasma	healthy	NaN	NaN	...	27.0	28	27	38.0	5.8	6.6	1.70	4.3	0.6	1.24
2	33	33	1_33_C8	BTO:0000131	NaN	NaN	blood plasma	healthy	NaN	NaN	...	18.0	21	69	18.0	6.2	5.7	1.12	4.1	0.5	1.12
3	34	34	1_34_C9	BTO:0000131	NaN	NaN	blood plasma	healthy	NaN	NaN	...	22.0	26	101	20.0	6.2	6.7	0.91	4.8	1.0	2.20
4	35	35	1_35_C10	BTO:0000131	NaN	NaN	blood plasma	healthy	NaN	NaN	...	18.0	25	61	13.0	5.4	5.5	1.21	3.9	0.4	0.90

5 rows × 46 columns

0. Import AlphaStats

Import library alphastats

import alphastats

1. Import Data

Load the MaxQuant proteinGroups.txt file and specify the columns indicating the intensity as well as the column that is used for indexing, like here the “Protein IDs” or the gene names. As the column is used for indexing, the values of this column must be unqiue.

maxquant_data = alphastats.MaxQuantLoader(file="../../testfiles/maxquant/proteinGroups.txt",
                                          intensity_column="LFQ intensity [sample]",
                                          index_column="Protein IDs")

2. Create a DataSet

Combine the imported MaxQuant data with the metadata

ds = alphastats.DataSet(
    loader = maxquant_data,
    metadata_path = "../../testfiles/maxquant/metadata.xlsx",
    sample_column = "sample" # specify the column that corresponds to the sample names in proteinGroups
)

AlphaStats will create a matrix of the Protein Intensities, which will be accessable using ds.mat and will save the metadata as a dataframe ds.metadata. Our original MaxQuant ProteinGroup file contains much more samples, than we have metadata for

## 3. Preprocess

print(f"Number of samples in the matrix: {ds.mat.shape[0]}, number of samples in metadata: {ds.metadata.shape[0]}.")

Number of samples in the matrix: 312, number of samples in metadata: 48.

Firstly, we will subset the matrix it will only contains samples, that are also described in the metadata.

ds.preprocess(subset=True)

print(f"Number of samples in the matrix: {ds.mat.shape[0]}, number of samples in metadata: {ds.metadata.shape[0]}.")

Number of samples in the matrix: 48, number of samples in metadata: 48.

Unnormalized data, Sample Distribution

sample_distribution_plot = ds.plot_sampledistribution(color = "disease")
iplot(sample_distribution_plot)

• Contaminations get removed indicated in following columns Only identified by site, Reverse, Potential contaminant (MaxQuant specific) and contamination_library (added by AlphaStats)

• Normalized using quantile normalization

• Missing Values get imputed using K-nearest neighbour imputation

ds.preprocess(
    remove_contaminations=True,
    normalization = "quantile",
    imputation = "knn"
)

/Users/drq441/opt/anaconda3/lib/python3.9/site-packages/sklearn/preprocessing/_data.py:2590: UserWarning:

n_quantiles (1000) is greater than the total number of samples (48). n_quantiles is set to n_samples.

After quantile normalization, Sample Distribution

sample_distribution_plot_2 = ds.plot_sampledistribution(method = "box", color = "disease")
iplot(sample_distribution_plot_2)

The preprocessing steps can be accessed using:

ds.preprocess_print_info()

Preprocessing:
The raw data contains 2611 Proteins/ProteinGroups.
The filtered data contains 2047 Proteins/ProteinGroups.Data has been normalized using quantile normalization.
Missing values were imputed using the k-Nearest Neighbor.
Contaminations indicated in following columns: ['Only identified by site', 'Reverse', 'Potential contaminant', 'contamination_library'] were removed. In total 202 observations have been removed.

4. Visualization

Principal Component Analysis (PCA)

pca_plot = ds.plot_pca(group = "disease", circle = True)
iplot(pca_plot)